My Gripes with this Game

I've been actively playing BeamNG.drive since late 2014 (around version 0.3.6.9) and have been a member of the forums since late 2015. I played it long before it was released on Steam and bought it quite early in development but had to wait until I had a PC that could run it. At first I really enjoyed the game, the updates were a bit more frequent a few years ago, and the game wasn't even at 0.4 yet. I thought it had plenty of time to continue adding stuff leading up to a mostly full release at 1.0. In August of 2016, I was satisfied enough with my first map to release it. Since then I have updated that map, remade Advanced Island, and am currently working on a very large highway map. We're at version 0.9 now, and we have a fair amount of maps, vehicles, and the beginnings of a campaign mode.

When I first started playing this game, my impression was that development would take awhile, and that was not a problem to me. The vehicle physics were superb and the maps were pleasant to drive through. I really hoped eventually the game would have some sort of campaign or career mode, and maybe races or rally stages against AI. Maybe even a multiplayer racing mode or an open-world map to drive around on with friends.

Nowadays, it seems as though the time between updates is a bit longer. That's okay, as long as they let us know what they're working on. But the devblog is rarely updated anymore, and all we have is the few posts in the development section each week. And this is an early access game, right? What happened to the experimental branch? Why don't you guys upload rougher versions of the game so we can test vehicle/map prototypes and report bugs? And what is truly being worked on?

From the looks of the development section and the most recent update, it looks like there are two maps being made right now. There was the new car, the Hopper, which is cool. But what else? We have a campaign mode, but what's the plan with that? What are the plans for the games? Why isn't there a development roadmap? Cars are cool and all, but the game still feels more like an engine demo. There is no real gameplay besides the campaign mode.

The AI is cool, but it's really not taken advantage of. Why can't we launch races against a few AI? Why aren't there rally scenarios where we race against an AI's time?

I mentioned that I make maps for the game, and the map editor is a whole other aspect of the game. There is a pretty big learning curve at first without much documentation as to where to start. Many users have created pretty good tutorials, and I'm fortunate enough to know some talented map creators. But the same bugs still plague the editor. And what happened to the scenario editor? I asked and wasn't given an answer. Nonetheless, the map editor is still a pretty well-made tool, but needs some work.

And performance. Normally I can run the game pretty well, I run it on high. I have an AMD FX-6300, 8 gb of RAM, and an 970 FTW 2.0+. I know the CPU is slightly outdated, but it gets the job done with almost everything I throw at it. But why is performance so different from map to map and vehicle to vehicle? Why do some map creators make maps that run better than official maps? Why isn't there documentation on how to optimize maps? I've been told this game really should need 16 gb of RAM, why haven't the recommended specs been updated to reflect that then?

Speaking of maps again, why isn't there a ton of support for larger maps? Myself and a few other map creators have created large maps and experience bugs. I know there are problems with the amount of objects that can be spawned and the performance of large maps.

And seriously, what are you guys working on? It's hard to tell from the lack of devblog updates and the occasionally unreleased map picture and car teaser. Why not have a weekly scheduled blog post detailing what has (or hasn't) been worked on? WIP content seems to be vaguely discussed and secretive. This is an early access game, the community is here to try the game in its incomplete stages.

I really enjoyed this game, and still do.You guys (the devs) are friendly and seem dedicated to the game. But I am concerned about the future of the game as it gets closer to version 1.0. I thought it would have many more features by this point. Let development take as long as needed. And if I've mentioned something that maybe has been addressed before, please let me know. I just wanted to write a post listing my concerns with the game right now in the hopes that someone else might feel the same way.

Kind regards,
-JD

 

I don't know, but for me there is huge amount of stuff they have worked on, but I guess problem is in this update most of the stuff was kind of invisible, some things that you don't really see or much even feel, but that take massive amount of work and when done they make much better possibilities for making new stuff you can actually see and experience.

In devblog there is huge list of changes, it says "3742 changes since last update, averaging around 47 changes per working day" that is features and improvements that allow bigger and better experience, so I would say they have worked hard, but things that has not been very visible.

From map to map difference in performance is something to do with object type, number of objects and most of the time related to shadows, but I'm too interested to know what is the key reason that would allow mod map makers to avoid performance pitfalls.

If you open that bigger more detailed FPS monitor with CTRL+F, you can see for example draw calls, lights etc. compare what happens when performance dies and when it runs smoothly. For example polycount seems not to matter much, drawcalls either, performance can die with 8000 draw calls as well as with 11000, but usually performance does not die not below 6000.

Reason FPS drop hard is because there is very high load on single thread, but why there is such load? Shadows mostly, but also something else.

Bob hinted that it would be draw call ceiling of DX11, but I'm not so sure about that, but what might be is that game engine has some CPU overhead per each drawcall and when you get enough many then performance dies as CPU delay increases.

However it can happen at 8000 draw calls or even 7000 in some cases, is it because there are other heavy stuff on same thread that sometimes performance dies earlier?

I have only questions, but I think Bob might be on right track there that draw calls might be related, however it is not quite as simple as that.

When FPS starts dropping, number of polygons and other data displayed by that advanced FPS counter start to jump around like crazy, so who knows maybe engine faces out of bounds situation? Again only guesses and questions, but something of that nature.

Shadows might be thing that pushes things over, without being actual reason behind it.

Visual distance is of course only real method that can be used to control this (tricks like using forest brush instead of other object types helps too I have read), but it should not affect terrain to be really useful, maybe that is area where devs look into improving performance in future, or maybe they are happy with current limits of island size and complexity, but I doubt that.

I can run 11 D15 at 60FPS now at pure grid map, but only one sunburst with rally lights at ECA town with high beams and fog lights on, second one drops FPS below 50FPS, I would think that dev team does have interest to improve this to be able to include some night time racing as physics is not limiting several cars, graphical side is.

Without knowing their plans all one can do is guess of course, so I guess if they would say something like withing next 12 months they seek to improve GFX performance might be what you would like to hear from them, but as they know what is limiting their goal they of course seek to improve that.

Now recently I remember reading they were looking C++ programmer for their team, if I go wild and make random guessing, I would think that might be related to them seeing need of improving Torque 3D's graphical performance and lifting limits it currently imposes of, but this is just random guessing.

On another thread it has been speculated that after 0.9 it might be 0.10.XXX versions, of course I have not seen any concrete evidence of that, but who knows, maybe?

Update: Bob found out that changing single non power of two texture to power of 2 dds improved minimum framerate from 30 to 45fps for him, that might be worth noting for sure.

 

I have a suggestion, do a thing like the beseige devs, a new post about development every 2 weeks. Not as in new vehicles, or maps, but as in game engine development. You could also combine the map progress announcements and car announcements all into one post every 2 weeks.

 

Yeah I really wish maps such as what is in ATS (the size and complexity) would be in BeamNG (ATS loads the map using sectors)

 

But what if they spent the last 2 weeks doing really boring stuff?

 

Still, lets us see that development is going on.

 

Someone has to do boring stuff too, it is no less important than fun and interesting stuff, but perhaps I would not like to read too much about boring stuff.

I kinda find that in modern world thing revolve too much around updates and new things, so maybe instead of new fixing something old would work as a contrast to all this modern crazyness. Maybe I need to grab a coat now

 

The AI is a bit too slow/oblivious for races to work well, the lack of support for large maps is due to engine limitations as has been explained plenty of times, and hopefully it will start getting a bit gamier soon. In one of the two recent updates they changed the value(as in price) of the race transmissions to be more realistic so that's a sign of them trying to balance a more grand campaign or career soon.

 

My only hat in the ring is, it does seem like the recent release has HUGE optimization improvements. First time I've actually been able to play with triple screens on my 960 and FX 8350. In one of the micro dev blog posts they mentioned deleting a few hundred thousand lines of code that was unnecessary. That's a hell of an improvement. I am actually getting around a 50% performance increase.
I really like these guys for not being like any of the other "sim" devs, being namely that they can take criticism, actually fix bugs, do it quickly, optimize like crazy, add new features/cars/maps... they do DO a lot, and from what I've seen scrubbing around the other sim racing eternal projects, these guys do a better job, in a cleaner fashion, quicker, and using a goofy little freeware engine with only like 6 people on the team.

I guess making a "gameplay" section of the game would be cool... but I'm just not sure what exactly it would be. I can't imagine it being more than the little scenarios we already have, but maybe with just some goofy cartoons like the campaign they put out. Not exactly the most enthralling thing in the world.

 

but who wants to write reports when you could be doing stuff?

 

That sounds like that your CPU has been limiting factor, I see 0% performance increase, but then again I rarely have more than 2 vehicles and physics were not limiting performance in 0.8 for me. Their maps are really nicely done and most of them manage to go around of game engine limitations so they run just fine, so I don't really need performance improvements, only game engine needs little bit of work to have less of those limitations, or at least map makers need information how to go around the problems best way.

However they do great job, lot more people can enjoy BeamNG now as you don't necessarily need that high end Intel to run the game, that mean of course much larger group of potential customers and more customers means better they can realize their dreams.

For me there really is no gripes, but like probably for most, more information and samples would be nice.

 

One thing is bugging me is what is the story with the Bluebuck? @gabester @B25Mitch give us something, even if it is a photo in a modeling program. please?
A pic of a stripped out shell in wcusa would be sweet.
like this.

 

My main issue right now isn't performance, but like I said, the lack of gameplay and progress updates. A progress update like NoxiousFumes suggested would be a good idea. I would also love to hear from one of the devs their opinion on anything I mentioned.

On another topic, I do not want to create maps for this game only to have them broken by the updates. I know that the game is in alpha, but when the game revolves around community creations, it is not a good idea to have them break mods.

 

Could be. I have a 8350 for two reasons, first, it's slightly cheaper (and I'm pretty sure I read DON'T PANIC in friendly letters on the box) and second, it is a bit faster in multicore rendering.

 

Really? The devs spent years fine tuning the collision physics, have gave every part and every car a price and recently adjusted the price of race transmissions, and added a save/load feature for damage. But you are thinking the most important part they will be lazy with and basically throw away all the work they did with balancing prices, working on AI, and adding the previously mentioned save/load feature? I'm thinking/hoping it would be more like SLRR mixed with the NFS games of the mid to late 2000's with some offroading and delivery missions.

 

No I really really hope they won't be lazy and throw away all the work. It's not laziness I'm worried about but more so taking on more than they can handle. They haven't given us any definite promises as to what kind of content the game is going to offer besides driving around maps and the campaigns. I'm just wondering if they don't have the ability to take the game to the next level. I would love some sort of online racing or rally, and delivery missions are a cool idea too.

 

The 0.9.x HOTFIX was actually partially because I went SPASTIC like a scalded house cat when 0.9.x update broke my map, and others, with multi-spawn for 0.7. I said something along the lines, "I worked 1000~1500 hours on my map, with an update due out the NEXT DAY, I'm gonna can this (Roane County Tennessee)", when they updated (the game) - vs updated when I was uploading it last November.
It's like hey "I worked so long on this, now it's belly-up and I can't even get into it, and I read the changelog and expected nothing like this". I agree, they should ADD features, but basic failsafes/regressive support for older content would be very helpful so at the very least, we can spawn in with a fly-cam and go from there, vs rendering content totally useless. You can't be running in circles bloating game code with everything, but a few actions when NIL is encountered, can go a long way without leaving too much bloat.
For them, in their corner, the argument is this is ALPHA, yes, and features are subject to change. Right, but if you're going to support modding, support it, don't burn your content creators, they're doing the same work you're making money doing, for free!
So I mean, to their credit, the dev's were very supportive, @synsol and @Nadeox1 (and in the past, @meywue), were very very supportive and helpful, in keeping mods working. THEY DON'T WANT TO BREAK THINGS, but there's no book that tells you exactly how to code everything to keep everything working, and no amount of people could test every mod and keep every one working while still having time to do all this, and the money to pay for it. To that end, that would take much time away from progress, and take features out of our hands; or atleast delay them greatly.

In the long run, as a mapper, I can CLEARLY feel the difference in the editor, it's a lot smoother, and it's also much, much, much more noticeable behind the wheel of any of the game's vehicles - you'll feel that in the first mile driven. I have put a car sideways at 60~65mph and never strayed from the lane I was in, in real life. I know how to make a proper emergency move in a vehicle, and haven't had an accident in over 13 years. This game is much, much more playable now then it's ever been before. It's really getting somewhere now, but there's always the price of entry, with everything in life.

On your behalf, @jdbensch , I agree, there should be CLEARER changes specifically for the content creators/modders on this site. We need to know what's changed under the hood so that, like me, we don't have to spend a day or two wondering why the @#$% (censored) the traffic light textures went out, or why the spawns crash the game, or why my jump texture suddenly does NOT work even though it's from the base-game, on a base-game model. On a 144 sq/mi map with 163 miles of roadway, it's a very arduous thing, to go and mess about with updating everything, and making sure you haven't missed anything.

We as modders understand the technical stuff that would absolutely boggle others and the laymen among us. We NEED that info. You can't develop content for anything when you don't know what you're running it on, in the end it means a lot of extra / double-work.

They are doing all the stuff under the hood, tying up all the loose ends, fixing all the deficiencies in the engine, and getting the full open-ended gameplay support in, so that we can have the game that was promised. Sure you want more features, but there is sooooo much to this. You might have to be older to understand, but for someone like me who's been in the modding community since it's inception in the DOS DOOM days, in the mid 90s there... I see the differences with every new version. Under the hood = not usually seen.
It may not look like much, but there's a lot to it, and in the end, this means a better end-result of a product. I can't even play GTA since this hit 0.4 or 0.5, GTA feels like a matchbox car game vs an awesome thing it was before Beamng.drive was out. I don't want a crime simulator because I am not a criminal acting out for attention, nor do I usually care for 'shoot em up' games. I like car driving, car customizing, guy stuff, typical, oh-so typical. I know how to make maps and I've wanted to make what I made / am making for the last 22~23 years since I started using pre-Pentium PC's (AM-486 dx/2 66 w/8mb RAM, 0.5GB Maxtor HDD, 2x CD, SB16, Trident 512kb VLB video, hey it was good then!)
So I started mapping for this game 16 months ago, and haven't looked back. It's always worth it, unless it can suddenly be taken away, then it's not worth it at all. So I totally see your point. Just keep mapping. If it doesn't work, contact support, or talk to me, usually I can figure most anything out, usually... No ones perfect!

Just be patient, as they always say "Good things come to those who wait". I can wait, and until then, I have plenty of pots on my stove to tend to, basically speaking, as a mapper, "A mapper's work is never done". That and "It's not over until it's over".
I shall recite the serenity prayer for those who don't know it, it helps keep me sane when nothing else seems to:
God grant me the ability to change the things I can change,
The patience to accept the things I cannot change,
and the WISDOM to know the difference.

I usually don't print things like that as it gets into religion/politics, and that's a touchy subject in general, but it's both helpful and relevant.

 

I don't see it that way. Making it something like a "life" sim could work, or again, some simple street races and you buy upgrade parts to progress. It's still pretty bare bones.
Beam has open world maps, which are quite small due to technical limitations... or so I assume anyways. They're fun for a while, and the Beam guys do a good job packing in a whole bunch of little back roads and things, but it also goes pretty quick. The most obvious thing for a game with an open world to do is... well... make it open world, but that doesn't work in Beams case. The physics can't handle a full open world with many things happening and AI randomly driving around. Even trying to do just a full race lineup can seriously bog down most peoples PC's.
The problem to me seems to be that for "gameplay" elements, Beam would have to stick to a simple street race type element. They have x amount of tracks, and only one of them is an actual racetrack, so theoretically you could work your way up to something like the Bolide supercar and do actual races, but that's really about it. Other than some old off road games (Typically in the PS2 era) I can't actually think of any open world driving games where the open world is actually empty. In those old off road games the only element of "career" is what I'm describing. Start at the bottom and buy your way up. It's not terribly exciting.
Other open world driving games have an actual world that at least pretends like somethings happening, and also usually a much larger environment. This includes everything from Forza Horizon, Burnout Paradise, the Driver series... to a small extent GTA (Since it's not ENTIRELY driving.) I simply can't see Beam accomplishing anything more than uncompelling PS2 era games. The big strive in the newer and more successful releases is having a world filled with things to do, and a clearly defined end goal and progression system. Beam doesn't have that. Unless they completely eliminate the free roam and make it mandatory to have to buy literally everything.

 

Simple street races were the kinds of things I was talking about where you buy your way up is basically what I meant, and is still how plenty of successful modern racing games like gran turismo work. Yeah the open world would be a bit empty without any other cars but it would serve its purpose. There is no reason that they would have to completely eliminate free roam, they could just have it as like a sandbox mode.

 

That's kind of where I'm concerned, because I agree with you. I just want to know the developers plans, whatever they might be. I'm still a little disappointed that I haven't gotten a response to this post from one of the developers, even just acknowledging they saw it.